Photocatalytic reduction of CO2 using surface-modified CdS photocatalysts in organic solvents

Journal

of’

Photoc&emistry Photobiology A:Chemistry

ELSEVIER

Journal of Photochclniatry

and Photobiology

A: Chemistry

I 13 ( 1998 ) 9347

Photocatalytic reduction of CO, using surface-modified CdS photocatalysts in organic solvents Bi-Jin Liu, Tsukasa Torimoto, Hiroshi Yoneyama * Accepted

2X October

1997

Abstract The photocatalytic reduction of carbon dioxide on CdS particles with and without surface modification wa:, investigated in various kinds of solvents.Forlmate and carbon monoxide were obtained as the major the former to the latter wax largely influenced by the kind of solvents when naked cadmium sulfide was increa.se of the dielectric constant of the solvent. l:he ratio increased. Even if CdS particles were modified the solvent was not eventually altered, but the ratio of formate to carbon monoxide became greater with of CdS.Theobserved effectof thesurfacemodification of CdS is discussed in terms of involvement of Cd’ of CdS photocatalysts in adsorption of anion radicals of CO,. 0 1998ElsevierScienceS.A. Ke~r~lc:

CO, reduction:

CdS photocatalyst:

Solvent

effect: Surface moditication

1. Introduction Photocatalytic reactionswith the useofsemliconductorpartitles have beeninvestigatedextensively 1I-61. The achievement of efficient photoinduced reductionsof carbon dioxide are the most desirableand challenginggoalsfor constructing artificial photosynthesissystems[ 7-261. Sofar, severalcompounds such as formate, carbon monoxide., methanol and methanehave beenreported asreduction compoundsofC0,. It hasalso been reported that changesof charged conditions of CdS photocatalyst surface 1141.adsorptionof In3+ metal ion on CdS photocatalyst surface [221, metal deposition on TiO, photocatalysts 19.17,I8 1, and changesof degreeof dispersion of photocatalysts in zeolite matricea [ 261 changed selectivity of CO? reduction. In our previous study using TiO, photocatalysts [ 271, it wasdiscoveredthat the polarity of solventsinfluencedgreatly on reduction routes of CO1 either to formate or to carbon monoxide, and the results were discussedin terms of differencesin adsorbability of CO,‘- anion radicals formed as a reaction intermediateon the photocatalyst surface.This paper deals with the photo-induced reduction of CO1 on CdS to confirm the validity of our discussionmade in the previous paper for TiO,. In addition to experimentsusing naked CdS photocatalysts. CdS having thiol compound:5as the surface /. Corresp<>nding

author.

IOIO-6O~O,‘9X/~lY.OO P//SIoIO-h03n(97)oo~lx-~

G lYY8 Elhevirr

by several kinds ofthiol compounds reduction products, and the ratio of used as the photocatalyst, and with by thiol compounds, the effect of increase of the surface modification sites adjacent to surface vacancies

Science S.A. All rights reserved

modifiers [ 28-371 was also usedto clarify roles of photocatalyst surfaceson the photoreduction of CO?.

2. Experimental section Cadmium sulfide of high purity (99.999%) was obtained from Mitsuwa Chemical Co. Ltd. These CdS particles had the hexagonal crystal structure as determinedby X-ray diffraction patterns, and had the average diameter of SOnm as determined by using a dynamic light scattering apparatus Otsuka ElectronicsDLS-70s Ar. Other chemicalsusedin this study were of reagemgrade and purchasedfrom Wako Pure Chemical Industry. Organic solvents were distilled prior to use.Water wasdistilled twice under atmosphericpressure. Five kinds of thiol compoundswere used in the surface modification of CdS after dissolving in appropriate solvents, that is, 0.1 mol dm ’ 2-aminoethanethiolin 2-propanol. 0. I and 0.5 mol dm -’ 1-dodecanethiol in 2-propanol, 0.1 mol dm-” sodium 2-mercaptoethanesulfonatein water. and 0.1 mol dm-’ sodium sulfide in water. CdS particles (0.72 g) were suspendedin these solutionsand the suspensionswere agitated overnight to achieve surface modification of CdS. By repeating centrifugation and washing with water, followed by drying in vacuum at 40°C for 12h, surface-modified CdS particles were prepared. The compositions of the

surface-modified CdS particles were determined by means of elemental analyses using a Perkin-Elmer. 240 C analyzerand atomic absorption spectroscopy using a Nippon Jarrel Ash, AA-8500 Mark II spectrophotometer. The photoreduction experiments of CO, were carried out using a quartz cell of a 7.0 cm3 capacity whose top was sealed with a rubber septum. The cell contained 3.0 cm’ of various kinds of solvents containing I .O mol dm--’ 2-propanol as a hole scavenger and 70 pmol CdS powder. After bubbling CO, for 30 min, the cell was closed and then irradiated with use of a 500 W high pressure mercury arc lamp as a light source and a 300 nm cut off filter. Irradiation intensity was I .O W cm- ’ for all cases. Carbon monoxide produced in the gas phase was determined using a gas chromatograph (Yanaco. C-2800). equipped with a molecular sieve 5A column (CL Sciences) and a thermal conductivity detector. Helium was used as a carrier gas. Acetone produced in the liquid phase was determined using a gas chromatograph ( Yanaco G- 180) equipped with a BX- 10 column (GL Sciences) and a flame ionization detector. A high-pressure liquid chromatograph (Tosoh CCPD) equipped with an organic acid column (waters) and a UV detector (Tosoh UV-800) were used to determine formate. The eluent used was 0.3% HJPOJ solution, and its flow rate was 0.80 cm’ min ’ The amount of samples taken from the cell for GC and HPLC analysis was 0.20 cm’ and 0.020 cm’. respectively. To measure fluorescence spectra, CdS particles were fixed to a glass plate ( I .O X 4.0 cm), and put in a quartz cell. Before measurements the cell was evacuated and then filled with Nz. Static fluorescence spectra of CdS particles were obtained using a Hitachi F-3010 fluorescence spectrophotometcr. Excitation wavelength was 220 nm.

Table I The molx analyses Catalyst

3. I. Cl~trrclcteri,latiorl

of CdS photocntulwts

Table 1 shows the molar ratio of adsorbed thiols to CdS determined by elemental analyses. and also shown in the table are surface coverage with thiol compounds which was determined by using the obtained molar ratio of thiol and the average particle diameter of 50 nm, and by assuming that the surface density of Cd’+ sites was the same as that of (000 I ) facet of hexagonal CdS. It has been reported that CdS particles emit fluorescence whose intensity is greatly influenced by adsorption of various chemical species [ 38421 such as OH- and S’~-. It is believed that surface vacancies of particle surface arc moditied by adsorption of these species. Fig. I shows fluorescence spectra of surface-modified CdS which arc given in Table I. In all cases, the shape of spectra are very similar to each other and a broad fluorescence peak which appeared at around 560 nm is assigned to fluorescence due to surface trap sites. However, the fluorescence intensity was a little different among

of adsorbed

modiiier

Surface-tnl)dilier”

to CdS

determined

by

elementitl

Ratio ol adaorhed

Surface coverage”

Fluolr\ce”ce intctlGty’

thml to CdS I I”lOl.‘i, I

I ‘/r )

(;L.“. I

II

“0°C

0

0

I .oo

h c d e

2-anlinoethanethiol I -dodecanethiol 2-mercaptoethaneaulfi~nate

0 60 0.60 0.90

O.Y3 0.77 0.72

sodium hultidr I -dodccancthiol’

“d” I.30

29 30 1s “d”

t

“See text for the concentration

ol moditicrs

0.67 0.63

65 and wlvents

wed.

“Calculated with use of ratio ot adsorbed thiol to CdS and particlc diameter. Details are drkhed in the text. ‘The relative intcnsity of Huorexxncr peak ut 560 nm k hhown in Fig. I. “Not determined. ‘03 mol dm ’ I -dodxanethiol

450

500

w;,> used

m wrt’xe

550

600

Wavelength Fig.

3. Results and discussion

wtio

I. The photoluminescencc

Notation!, wit’. 220

of I a) to ( f) tire those nm.

spectra fivett

modttication

of CdS.

650

700

of surface-modified

CdS

/ nm in Table

I, The cxcitatwn

pxtlcleh.

wuwlenfth

the CdS particles used. If it is assumed that a decrease of the amount of surface trap sites causes decrease of the intensity of the fluorescence. the results given in Fig. I suggest that the surface state density decreased from (a) to (f). This suggestion accorded qualitatively with the degree of surface moditication as shown in Table I,

As shown in Fig. 2, irradiation of naked CdS particles in CO,-saturated acetonitrile and dichloromethane containing 2-propanol as a hole scavenger resulted in the formation of formate and/or carbon monoxide as the reduction products of CO, and simultaneous production of HI. As the counterpart reaction of these, acetone was produced from 2-propanol. and no other oxidation and reduction products were obtained. The amount of each product increased linearly with time. indicating that activities of the photocatalyst were not decreased

100

*_--

r-

0

II 0

2

4 Time

6

8

0

2

4 Time

Ih

Fig. 2. The time course of thr production of formate (0 c 3). hydrogen ( v ). and acetone ( 4 ) from (‘Ol-aaturatcd and dichlorcmwthane (b) containing tocatalyst u\ed was the naked CdS.

I .O mol dm

6

0

8

1, carbon monoxide acetonitrile (aJ

’ ?-propanol.

The pho-

CO,+e~~--+CO~W(ad)

(I)

CO>~(ad)+Hf+e~+HCOO~

(2)

CO> +CO+e~

+CO+OH-,CO+COf-

40

60

80

Dielectricconstant

during the photoreduction experiments. It was found by comparing the sum of the amount of reduction products with that of the oxidation products (acetone) that the chemical stoichiometry ofthe reduction andoxidation products were maintained. Furthermore it is noticed that the kinds of solvents influenced greatly reduction behaviors of CO2 as observed in the use of TiO, photocatalysts 127 1; CO, reduction in acetonitrile gave carbon monoxide and thrmate, while the carbon monoxide was selectively produced in dichloromethane. In order to investigate the effect of solvents on reduction behaviors of CO2 in more details. photoreduction expcriments were carried out using the naked CdS in various solutions. It was found that in all cases, formate and/or CO were produced and no other reduction products ‘were detected. Fig. 3 shows the fraction offormate in CO2 reduction products as a function of the dielectric constant [ 333.51 of solvents used. In this figure, the results obtained with the use of QTiO,/SiO, which were reported in our previous paper [ 37 ] are also included. In both cases, the fracti’on of formate increased with increase of the dielectric constant of the hoIvents used, though the selectivity of the reduction reactions was a little different between CdS and Ti02, that is. CdS has a tendency to produce more CO than Ti02. As already discussed in the previous paper 1271. photocatalytic reduction of CO, may be formulated by Eqs. ( 1) and (2 )Eqs. (3a) and (3b). where protons were produced by oxidation of 2-propanol with photogenerated holes.

CO’,-(ad)+H’+e

20

ih Fig. 3. The f’rtaction

ofthrmate

\oIvent wed. The rewlts photocatalyst for 7 h ( n talyst ( 0) arc also shown.

a~ a function

were obtained ). The results The solution

ofthe

d&ctric

constant

hy irradiationof the withurc ot Q-TiO,/SiO, contained

lor all cilscs. Solvent used \bete a. carbon c, ?-propnnol: d. propionttrk: e. ethylene mile; -c. ~ulfolane: h. propvlenc carbonate: i

I .O mol dm

ofthc

naked CdS photoca’ 2.propanol

tztrachloride: b, dichloromethanr: ylycol monoethyl ether: f. xctonand i. war.

anion radicals. becausethe CO,’ anion radicals are not highly solvated in solvents of low polarity. If this is the case, carbon monoxide will be produced as the major reduction product of CO,. asdiscussedpreviously. If solvents of high dielectric constantsuchaswater and propylene carbonateare used,the formed CO?‘- anion radicals may be greatly stabilized by the solvents, resulting in weak interaction with photocatalyst surfaces. Then the carbon atom of CO,‘- anion radicalstendsto react with a proton to give formate. The fact that the carbon monoxide production was more favorable in the case of using CdS particles rather than Q-TiO,/SiOl seemsto suggestthat CO?‘~ anion radicals tend to adsorb more easily on the CdS surface rather than Q-TiO,/SiO, surface. Fig. 4 shows time course of photoreduction products ot CO, in dichloromethanewith useof I -dodecanethiol-moditied CdS. The surface coverage of the thiol modifier wasca. 65%. Being different from the resultsobtained in the useof the nakedCdS photocatalysts,hydrogen production occurred predominantly asreduction reaction, and as for photoreduction of CO,. both formate andCO were formed while acetone

(Xl) (3b)

In our previous paper [ 271. we discussed that adsorbability of co,= on photocatalyst surfaces determines which one. Eq. ( 2) or Eqs. ( 3a) and ( 3b 1, occurs more predominantly. If low polar solvents such as Ccl, and CH2CI, are used, the formed CCLP anion radicals may be adsorbed strongly on Cd sites ofCdS surface through the carbon atom of CO?‘-

0

4

2 Time

6

8

I h

Fig. 1. The time comx of the production of fol-mate (U), carbon monoxide IO). hydrogen C v ), and ticetonc I + ) from CO?-saturated dichloromrthane wlution containing 1.0 mol dm- ’ 7.propanol. The photocatalyst ured was the I-dodecanethiol-moditicd CdS particle (hampIe (f) in Table I ).

96

coverage of the modifier. As already discussed above. the fluorescence intensity is correlated to the abundance of sulfur vacancies, and CdS photocatalysts which emit high fluorescence contain a lot of sulfur vacancies. The net charges ofCd sites on the particle surfaces seem to be influenced by the abundance of the sulfur vacancies, and the higher the amount of sulfur vacancies. the greater the surface charges of Cd sites.

100 tj a

80

g z

60

Ei G 0 h

40 20 0

b

c

d

100 t; 4 e g .=s P h

(b)

References

80

I I ] A. Henglcin. Top. Curl-. Chrm. l-13 ( 1988) I I3 [ 21 P.V. Kamat. Chem. Rev. Y3 ( 1093) 267. 131 M.A. Fox.M.1‘. Dulay.Chem. Rev. 93 ( lYY3) 331.

60 40

1-l) H. Yoneyama, Cl-it. Rev. Solid State Mater. Sci. IX I 1093) 6Y. 15 I A. Hql’cldt. M. Gr&el\alignl ch. Chem. Rev. Y5 ( IYYS) 19. [ 61 A.J. Nwik. R. Mrmming. J. Phys. Chem. 100 ( IYY6 I 13061.

20 0

a

b

c

d

e

f

Photocatalyst Fig. 5 The fraction of CO, reduction fxoduct\ ohtoinrd by irradiation ot vwious wrf’xe-modified CdS photncataly\t\ for 7 h in wxtnnitrilc (A ). and dichloromrthane (B 1 containmg I .O mol dm ’ ?-prnpxwl. Photocatalyst\ of (a I to ( t) were the wme as thee giLen in Table I.

was produced as an oxidation product. Since a good chemical stoichiometry was obtained between the reduction and oxidation products, it seems unlikely that photodestruction of thiol compounds adsorbed on CdS particles contributed to the obtained results. Fig. 5 shows the fraction of formate and CO in CO2 reduction products obtained for use of various kinds of CdS particles in acetonitrile and dichloromethane. Hydrogen was excluded in the calculation of the fraction. It is clearly seen that the fraction of formate to CO varied greatly depending on the kind of CdS photocatalysts used. and in both solutions. it increased in the order of CdS photocatalysts from (a) to ( f). Though the same moditier of I -dodccanethiol was used in the photocatalysts (c) and ( f). the fraction of CO1 reduction products obtained was greatly different. Referencing the results shown in Table I for fluorescence intensities, the fraction of formate production (or that of CO production) seems to be related to the activ,ities for the fluorescence emission. The presence of functional groups of ~ NH2 and SO, in thiol compounds did not greatly alter this tendency. How the fraction of formate production is correlated to the fluorescence emission is then a matter for discussion. It seems likely that CO,‘- anion radicals in low polar solvents tend to adsorb on the positively charged Cd’+ sites. resulting in the CO formation. Since the surface moditiers used in this study should be bound to surface Cd” sites, the amount of adsorbed CO,‘~- on Cd’ ‘~ sites must decrease with increase of the degree of surface modification. and then the formate production becomes predominant with increasing the surface

17 I T. Inour. A. Fujlshmxr. S. Konishi. 637. I XI A. Henglein. M. GutierreL. C.H. Chcm. XX I IYXJ) 170. [Y I K. Trnnakonc. Solar Energy Mater. I I01

B.R. EfGns. J.T.S. Irvnw. Chem. Commun., ( IYXX) 1 I I 1 S.M. Aliwi. K.F. Al-Juhori.

K. Honda. Fischer.

Nature Brr.

10 ( 1984)

E.P. Murphy. 1113. Sol. Energy

177

f 1079)

Bunacngeh.

Phyh.

235.

J. Grimshaw. Mater.

J. Chem.

Id ( 19x9)

Sot..

713.

j 121 J.T.S. II-vine. B.R. &$I~\. J. Grimshaw, J. Sol. Energy ( IYYO) 17. I 13 1 H. Inow, T. Torimoto. T. Sakata. H. Mori, H. Yoneyama. Chem. Lett. I IYYO) 13x3. I I4 I H. Inouc. R. NJkamurn, H. Yoneyama. Chem. Lett. ( 199-I) 1227. 1 IS I S. Kuwabnta. H. Uchida. SW., Chcm. Commun. I 161 H. Inoue. T. M,ltwy;lma.

A. Ognwa. S. Hirao. ( IYYS ) X29. B.-J. Liu. T. Sakata.

H. Yoneyama. H. Mori.

J. Chrm.

H. Yoney;lma.

Chcm. Lrtt. ( I YY4) 653. 1 I71 P. Alhrl-. J. KIWI. New J. Chem. I3 ( IYYO) 135. I IX] Z. Gwen. I. Willnct-. A.J. Nelwn, A.J. Frank. J. Phys. Chem. Y4 ( IYYO) 3784. 1 I Y 1 M. Knnrmoto. ‘T. Shirq;m~i. C. Pac. S. Yanagida. Chem. Lett. ( 1091)) 931. [ 101

M. K~nrmoto. gida. Chem. [ 2 I I 41. Kancmoto. (1992) 3521. [ 27 1 M. Kunemoto. Lett. t lYY3)

K.-I. Ishiharu. Y. Wada. T. Sahata, Lrtt. ( lYY2 I X.35. T. Shmagami, C. Pac. S. Yanagida. M. Nomura. 10x7.

Y. Wada.

[ 23 1 Y. Matwmoto. M. Ohata. J. Hombo. ] 341 M. Anpo. K. C‘hiha, J. Molar. Catal. j 25 I M. Anpo, H. Ylmashita, Y. Ichihashi. 3Yh (IYYS) M. Anpo, Chem.. B 27 I B.-J. LIU. Photobitrl.

[ 261

T. Akano.

H. Mot-i. J. Phys.

S. YanaChem.

S. Yanagida.

96

Chem.

J. Phys. Chem. 98 ( IYY4) 2YSO. 74 I IYY:!) 207. S. Ehara. J. F.lcctroanal. Chrm.

21 H. \r ama\hlta. Y. lchihashi. IO1 t IYY7) 2631. T. Twimoto. H. Mntaumoto. A: C’hcm. I OX I I YY7) I X7.

Y. FuJii.

M. Honda.

H. Yoneyamn.

J. Phyb.

J. Photrxhem.

2X I Y. Nosaka. K. Yamaguchi. H. Miyuma. H. Huyashi. Chcln. Letts. ( IYXX) 60.5. 291 M.L. Stclgerwald. A.P. Al~v~aatos, J.M. Gibbon. T.D. Harris. R. Kortan. A.J. Muller, A.M. Thayer. T.M. Duncan. D.C. Dougla\\, L.E. Brus. J. Am. Chem. Sot. I IO f I YXX) 3046. 301 D. Haye>. 0.1. Micic, M.T. Ncnadovic. V. Swayambunnthan, D. Mcivzl. J. Phy\ Chrm 93 t IYXY) 1603. 31 1 C.H. Fiwher, A. Henglein. J. Phyh. Chem. Y3 ( 19X9) 5578. 32 I V. Swaydmhunuthan. D. Haye\. K.H. Schmidt. Y.X. Line, D. Meisel. J. Am Chum. Sot. I I? ( IYYO) 3X3 I, 33 ] N. Hel-ran, Y. Wang. H. F.chert. J. Am. (‘hem.

Sot.

I 12 ( IYYO)

1.322.

33 1 V.I.. Calvin. A.N. Goldswin. I I YY2 ) 512 I. 3.5 I .A. Chemscddine. H. Weller,

A.P. Aliviwwr. Brr.

Bunwngc\.

J. Am. Chrm. Phy<. C‘hem

SIC.

I II

Y7 ( lYY3)

h36. 36 1 T. Torimtrto. K. .Marda. J. Macnaka. H. Yonryama. J. Phys. Chrm. YX ( IYW) l.%l58. 37 1 M Miy.lkc. H. Ma~wmo~. M. Nid~i/awa, T. Sahara. H. Mori, S. Kuu,~hn~n. H. Yoneyama. Lanpuir 1.2 ( IYY7 1 7i.7. 38 1 A Hcnglein. Ber. Bunrcnpe\. Phy\. Chem. Xh ( 19X? 3Y j T. II;mnhauxr. M. O’Ncil. K. Joh;n\\on. D Whitten. J. 1%) \. Chcm. YO ( I ‘JXh) hO7-I.

I 301, G. Mclendton.

1%) I L. Spanhel. M. Haaw. H. Wcllcr. IO’) llYX7) 5649. 1-II 1 M. O‘Ncll, J. Marohn. G. Mclendon. 142 I U. Rcxh, H.M. Eyctlmlillel-vslifnl x I IYY?) 2215

A. Henglem.

J. Am.

J. Phys. Chrm. ch. H. Wellrrvalign

Chem.

Sot.

Y-t ( IYYO) 13%. I ch. Laymuir

I33 I The Chemd Strclety of Jalun (13th. ). Kagaku Binran. Vd. I. Marufxm Inc.. IYX1. pp. Sh I, 144 j R.C. Wcat ( Ed. 1. Handhook 01 Chcmlwy and PhyGc\. CRC PVC\\ Inc.. IYhY, p, E-61